In order to study the expression of genes underlying physiological development of the CNS, we have developed a quantitative PCR technique for the enumeration of mRNA molecules. Over the course of rat cervical spinal cord development we have measured the expression of the GAD (glutamic acid decarboxylase) gene family. GAD, GABA receptors and neurofilament mRNAs (as a neuronal marker) appeared in parallel, suggesting that GAD plays a role in neurogenesis. The precisely quantitated time course of GAD was described by a mathematical model based on feedback regulation of GAD on its own expression. This model provides a working hypothesis, making experimentally testable predictions. Furthermore, an alternatively spliced transcript of GAD67, EP10, was most strongly, almost restrictively expressed during development and is implicated in the early functioning of GAD in the transition from proliferation to differentiation. The parallel decline of EP10 and nestin (a neuroprogenitor marker) and number of proliferating cells are in accord with this conjecture. In order to identify other developmentally important genes, we have conducted a quantitative PCR survey of genes associated with intra~ and intercellular signaling in the cervical spinal cord. Genes coding for neurotransmitter synthesizing enzymes and receptors, peptide factors and their receptors, and genes associated with neurogenesis were differentially expressed during development. In addition, members of the IP3 (inositol trisphosphate) receptor family, playing a central role in regulating the second messenger, Ca2+, were distinctly regulated during neurodevelopment on the level of mRNA expression. We have established a new method based on arbitrarily-primed PCR enabling the parallel identification of hundreds of randomly selected expressed genes. Using this technique, we have compared RNA samples from NGF-treated and naive PC12 cells and E12, E18 and adult cervical spinal cord, and have identified on the order of 30 to 50 differentially expressed genes. The amplified cDNA fragments have been isolated, sequenced and compared to known genes in the GenBank and EMBL databases. Several differentially expressed sequences matched with known genes: L1 line element (a retrotransposon), ribosomal protein S3A, cytochrome C oxidase and Zn/Cu- dependent superoxide dismutase. Some of the genes identified here have also been detected by subtractive hybridization of growth-factor treated cell cultures. The results confirm that the method faithfully detects differentially expressed genes and is suitable for the compilation of unbiased gene expression spectra.